ef29cd9bde8695645d63ae7f568fb298809e7b19,tpot/builtins/dataset_selector.py,DatasetSelector,fit,#DatasetSelector#,43

Before Change

        features_i_df = pd.read_csv(self.subset_i, sep="\t", header=0)
        if isinstance(X, pd.DataFrame): // use columns" names
            self.feature_names = list(X.columns.values)
            feature_i = [str(val) for val in features_i_df.values.flatten()]
        elif isinstance(X, np.ndarray): // use index
            self.feature_names = list(range(X.shape[1]))
            feature_i = [int(val) for val in features_i_df.values.flatten()]

After Change

        
        subset_df = pd.read_csv(self.subset_list, header=0, index_col=0)

        if isinstance(self.sel_subset, int):
            self.sel_subset_name = subset_df.index[self.sel_subset]
        elif isinstance(self.sel_subset, list):
            self.sel_subset_name = []
            for s in self.sel_subset:
                if isinstance(s, int):
                    self.sel_subset_name.append(subset_df.index[s])
                else:
                    self.sel_subset_name.append(s)
        else: // self.sel_subset is a string
            self.sel_subset_name = self.sel_subset

        sel_features = subset_df.loc[self.sel_subset_name, "Features"]
        if not isinstance(sel_features, str):
            sel_features = ";".join(sel_features.tolist())

Instances